From 5cbd37e5056f6a715333553e9d529e0ddddb3e78 Mon Sep 17 00:00:00 2001
From: Dan Gohman <gohman@apple.com>
Date: Thu, 6 Aug 2009 09:18:59 +0000
Subject: [PATCH] Fix a few places in DAGCombiner that were creating
 all-ones-bits and high-bits values in ways that weren't correct for integer
 types wider than 64 bits. This fixes a miscompile in PPMacroExpansion.cpp in
 clang on x86-64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78295 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/CodeGen/SelectionDAG/DAGCombiner.cpp | 22 +++++++++++++++++-----
 test/CodeGen/X86/wide-integer-fold.ll    | 12 ++++++++++++
 2 files changed, 29 insertions(+), 5 deletions(-)
 create mode 100644 test/CodeGen/X86/wide-integer-fold.ll

diff --git a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 7f2c91bc5c0..3d0f337c709 100644
--- a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -1930,7 +1930,7 @@ SDValue DAGCombiner::visitOR(SDNode *N) {
 
   // fold (or x, undef) -> -1
   if (N0.getOpcode() == ISD::UNDEF || N1.getOpcode() == ISD::UNDEF)
-    return DAG.getConstant(~0ULL, VT);
+    return DAG.getConstant(APInt::getAllOnesValue(VT.getSizeInBits()), VT);
   // fold (or c1, c2) -> c1|c2
   if (N0C && N1C)
     return DAG.FoldConstantArithmetic(ISD::OR, VT, N0C, N1C);
@@ -2477,9 +2477,13 @@ SDValue DAGCombiner::visitSHL(SDNode *N) {
     uint64_t c1 = cast<ConstantSDNode>(N0.getOperand(1))->getZExtValue();
     if (c1 < VT.getSizeInBits()) {
       uint64_t c2 = N1C->getZExtValue();
+      SDValue HiBitsMask =
+        DAG.getConstant(APInt::getHighBitsSet(VT.getSizeInBits(),
+                                              VT.getSizeInBits() - c1),
+                        VT);
       SDValue Mask = DAG.getNode(ISD::AND, N0.getDebugLoc(), VT,
                                  N0.getOperand(0),
-                                 DAG.getConstant(~0ULL << c1, VT));
+                                 HiBitsMask);
       if (c2 > c1)
         return DAG.getNode(ISD::SHL, N->getDebugLoc(), VT, Mask,
                            DAG.getConstant(c2-c1, N1.getValueType()));
@@ -2489,9 +2493,15 @@ SDValue DAGCombiner::visitSHL(SDNode *N) {
     }
   }
   // fold (shl (sra x, c1), c1) -> (and x, (shl -1, c1))
-  if (N1C && N0.getOpcode() == ISD::SRA && N1 == N0.getOperand(1))
+  if (N1C && N0.getOpcode() == ISD::SRA && N1 == N0.getOperand(1)) {
+    SDValue HiBitsMask =
+      DAG.getConstant(APInt::getHighBitsSet(VT.getSizeInBits(),
+                                            VT.getSizeInBits() -
+                                              N1C->getZExtValue()),
+                      VT);
     return DAG.getNode(ISD::AND, N->getDebugLoc(), VT, N0.getOperand(0),
-                       DAG.getConstant(~0ULL << N1C->getZExtValue(), VT));
+                       HiBitsMask);
+  }
 
   return N1C ? visitShiftByConstant(N, N1C->getZExtValue()) : SDValue();
 }
@@ -3094,9 +3104,11 @@ SDValue DAGCombiner::visitSIGN_EXTEND(SDNode *N) {
     }
     
     // sext(setcc x, y, cc) -> (select_cc x, y, -1, 0, cc)
+    SDValue NegOne =
+      DAG.getConstant(APInt::getAllOnesValue(VT.getSizeInBits()), VT);
     SDValue SCC =
       SimplifySelectCC(N->getDebugLoc(), N0.getOperand(0), N0.getOperand(1),
-                       DAG.getConstant(~0ULL, VT), DAG.getConstant(0, VT),
+                       NegOne, DAG.getConstant(0, VT),
                        cast<CondCodeSDNode>(N0.getOperand(2))->get(), true);
     if (SCC.getNode()) return SCC;
   }
diff --git a/test/CodeGen/X86/wide-integer-fold.ll b/test/CodeGen/X86/wide-integer-fold.ll
new file mode 100644
index 00000000000..64c6d98616b
--- /dev/null
+++ b/test/CodeGen/X86/wide-integer-fold.ll
@@ -0,0 +1,12 @@
+; RUN: llvm-as < %s | llc -march=x86-64 | FileCheck %s
+; CHECK:  movq  $-65535, %rax
+
+; DAGCombiner should fold this to a simple constant.
+
+define i64 @foo(i192 %a) nounwind {
+  %t = or i192 %a, -22300404916163702203072254898040925442801665
+  %s = and i192 %t, -22300404916163702203072254898040929737768960
+  %u = lshr i192 %s, 128
+  %v = trunc i192 %u to i64
+  ret i64 %v
+}
-- 
2.34.1